In the semiconductor processing industry, ions are often implanted into a workpiece, such as a semiconductor wafer, in order to provide specific characteristics in the workpiece. One common process involves implanting ions into a workpiece, wherein transistor devices have been previously formed and isolated across the workpiece, and wherein a polysilicon contact is positioned over a gate of the device. The gate further overlays a so-called “well”, wherein contacts to the well generally define source and drain contacts for the device, therein defining terminals of the device. A thin oxide further resides between the gate and the channel, wherein the contacts on either side of the gate define the source and drain. In operation, when a positive voltage is applied to the gate, such as in an n-channel transistor device, the positive voltage enhances or attracts negative charge and pushes out the positive charge, therein effectively cutting off conduction through the transistor. When the positive voltage is relaxed, charge is allowed to enter into the channel, therein permitting the transistor device to conduct.
When a positive voltage is applied to the gate that exceeds specifications for the oxide (e.g., a voltage that would create a relatively high electric field in the oxide on the order of 5-10 MV/cm), a current will generally start to flow through the oxide. Initially, current flows through the oxide via quantum-mechanical tunneling current (often referred to as Fowler-Nordheim or FN tunneling current) or direct tunneling, and the initial current flow typically produces no significant damage to the thin oxide, as little to no heat is initially produced during the initial flow of current. Over time, however, charge traps are generated by the current flow, thus eventually causing the oxide fail. A relatively large amount of charge (e.g., 1-3 coulombs/cm2) is typically required to flow through the oxide before the oxide breaks down or fails.
The voltage at which current is initially conducted in a known, good, oxide is quite predictable. For example, for a given oxide thickness, the tunneling current is typically known, and can start at around 6-10V. During semiconductor processing, such as during an ion implantation process, it is desirable to determine whether the ion implantation will cause the device to reach the tunneling voltage, and if it does, whether current flow exists.
Conventionally, charge monitors have been utilized to measure the peak voltage that is impressed on the workpiece by an ion beam or ion implantation process. Such a peak voltage is commonly measured using a floating probe (e.g., a Langmuir probe). The floating probe is typically a planar probe (e.g., a small disc, approximately 1 mm in diameter), wherein when the probe is passed through the ion beam, it experiences either a positive or negative charging voltage, depending on whether there is an excess of ions or electrons in the beam. Typical charging voltages are in the range of +/−10V. In the small devices implemented in modern semiconductor processing, however, such charging voltages are often enough to induce a current to flow within the gate oxide structures.
Such a charging voltage inducing a current flow, however, does not necessarily indicate, by itself, that damage and/or break down is occurring within the device. Further, such a peak voltage measurement fails to provide enough information to determine whether damage to the device occurs, because as a general rule, the tunneling current is not damaging to the workpiece.
Another conventional methodology to monitor charges has been the use of consumable monitor wafers. Monitor wafers (also called test element group wafers) are comprised of semiconductor wafers having various capacitor structures formed thereon, wherein the capacitor structures have large contacts coupled thereto. The large areas of the contacts collect a relatively large amount of charge and focus it on a small capacitor gate. Various sizes of contact areas and gate areas for the gate contacts (e.g., the thin oxide), are provided, with the remainder of the device residing over a thick oxide layer. When a voltage is impressed across the monitor wafer, the flow of current in the device is focused onto the gate, itself. Various ratios of areas of contacts to areas of gates (called the “antenna ratio”) are provided, and give a large current density in the gate, itself, such that the failure of the monitor devices is accelerated. Monitor wafers, however, are very expensive, and are used as a consumable or disposable test wafer for an implant.
Accordingly, a need exists for a new, more robust, and inexpensive bipolar measurement system and methodology for in-situ measuring of charge build-up during ion implantation.